data.isnull().sum()

plt.figure(figsize=(7,4))

sns.heatmap(data.isnull())

data[data["Age"]<=0].shape

data["PassengerId"].duplicated().sum()

numerical_data = []

object_data = []

​

for column in data.columns:

    if data.dtypes[column] != 'object':

        numerical_data.append(column)

    else:

        object_data.append(column)

numerical_data

from sklearn. impute import KNNImputer

imputer = KNNImputer(n_neighbors=5)

data[numerical_data] = imputer.fit_transform(data[numerical_data])

data.isnull().sum()

for column in data.columns:

    missing_indices = data[data[column].isnull()].index

    available_values = data[column].dropna()

    for index in missing_indices:

        random_choice = np.random.choice(available_values)

        data.at[index, column] = random_choice

data.isnull().sum()

data['Fare']=data['Fare'].round(2)

plt.figure(figsize=(7,4))

sns.heatmap(data.isnull())

plt.figure(figsize=(7,5))

px.box(data_frame=data, y="Fare")

sns.distplot(data["Fare"])

plt.show()

#find the limits

upper_limit=data["Fare"].mean() + 3*data["Fare"].std()

lower_limit=data["Fare"].mean() - 3*data["Fare"].std()

print("upper limit: ",upper_limit)

print("lower limit: ",lower_limit)

#find the outliers

outliers_df=data.loc[(data["Fare"]> upper_limit) |(data["Fare"] < lower_limit)]

outliers_df.shape

plt.figure(figsize=(7,5))

px.box(data_frame=new_df, y="Fare")

sns.distplot(new_df["Fare"])

plt.show()

file_path="new df.csv"

num_columns=list(data.drop(['Pclass','Survived'],axis=1).select_dtypes(include=['float','int']).columns)

num=int(len(num_columns)/2) if int(len(num_columns)/2)>1 else 2

fig ,ax = plt.subplots(num,num,figsize=(12,10))

for j in range(num):

    for i in range(num

):

        try:

            sns.histplot(data=df,x=num_columns[0],kde=True,bins=20,ax=ax[j][i])

            num_columns.pop(0)

        except:

            fig.delaxes(ax=ax[j][i])

fig.suptitle('Histograms of numerical columns', fontsize=16)

plt.show()

cat_columns=list(data.select_dtypes(include=['object']).columns)

num=int(len(cat_columns)/2) if int(len(cat_columns)/2)>1 else 2

fig ,ax = plt.subplots(num,num,figsize=(12,10))

for j in range(num):

    for i in range(num

):

        try:

            sns.countplot(data=df,x=cat_columns[0],ax=ax[j][i],hue='Survived')

            cat_columns.pop(0)

        except:

            fig.delaxes(ax=ax[j][i])

fig.suptitle('Histograms of numerical columns', fontsize=16)

plt.show()

num_columns=list(data.drop(['Pclass','Survived'],axis=1).select_dtypes(include=['float','int']).columns)

num=int(len(num_columns)/2) if int(len(num_columns)/2)>1 else 2

fig ,ax = plt.subplots(num,num,figsize=(12,10))

for j in range(num):

    for i in range(num

):

        try:

            sns.histplot(data=df,x=num_columns[0],kde=True,bins=20,ax=ax[j][i])

            num_columns.pop(0)

        except:

            fig.delaxes(ax=ax[j][i])

fig.suptitle('Histograms of numerical columns', fontsize=16)

plt.show()

num_columns=list(data.drop(['Pclass','Survived'],axis=1).select_dtypes(include=['float','int']).columns)

num=int(len(num_columns)/2) if int(len(num_columns)/2)>1 else 2

fig ,ax = plt.subplots(num,num,figsize=(12,10))

for j in range(num):

    for i in range(num

):

        try:

            sns.boxplot(data=df,x=num_columns[0],ax=ax[j][i])

            num_columns.pop(0)

        except:

            fig.delaxes(ax=ax[j][i])

fig.suptitle('Histograms of numerical columns', fontsize=16)

plt.show()

import scipy.stats as stats

from sklearn import metrics

from sklearn.metrics import classification_report

%tensorflow_version 1.x

import tensflow as tf

target=data.Survived

age_groups = pd.cut(data['Age'], bins=range(0, 100, 10), include_lowest=True)

pd.crosstab(age_groups,data.Survived).plot(kind='bar')

plt.ylabel('No.of Passengers')

plt.title('Age over Survival')

plt.grid(color="red", linestyle=":", alpha=0.5)

pd.crosstab(data.Sex,data.Survived).plot(kind='bar')

plt.ylabel('No.of Passengers')

plt.title('Sex over Survival')

plt.grid(color="red", linestyle=":", alpha=0.5)

fare_groups = pd.cut(data['Age'], bins=range(0, 550, 10), include_lowest=True)

pd.crosstab(fare_groups,data.Survived).plot(kind='bar')

plt.ylabel('No.of Passengers')

plt.title('Fare over Survival')

plt.grid(color="red", linestyle=":", alpha=0.5)

from sklearn import tree

model = tree.DecisionTreeClassifier(criterion='entropy', max_depth=3)

from sklearn import tree

model = tree.DecisionTreeClassifier(criterion='entropy', max_depth=3)

inputs=data[['Pclass','Sex','Age','Fare']]

inputs

from sklearn.preprocessing import LabelEncoder

obj=LabelEncoder()

inputs['Sex']=obj.fit_transform(inputs['Sex'])

inputs.head(4)

from sklearn.model_selection import train_test_split

x_train, x_test, y_train, y_test = train_test_split(inputs,target, test_size=0.3)

model.fit(x_train,y_train)

model.score(x_test,y_test)

model.predict([[2,1,22,100]])

model.predict_proba([[2,1,22,100]])

​